Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Need Help In Fiter Designing Of Three Phase Inverter...:(

Status
Not open for further replies.
Hi,

I am not sure what you want to know. The output would ramp up from some low value while the frequency stayed constant, until the point where the voltage got up to the normal voltage/frequency relationship and then it could either change both or just stay there.
So lets say your motor works at 20v and 20Hz. You might start at 20Hz and start the output voltage at 5v and ramp the output voltage up little by little. You should not have to change the input DC voltage.
 
Thank you for the info, for snubbing why don't you just have fast diodes across each device? As long as the DC supply is well decoupled by metal film caps near the devices that should deal with your ringing. But it is very layout dependent.
The common core for your supplies shouldn't be a problem since the windings are isolated. But a spike in one supply will typically be coupled to all.
Soft start is to ramp voltage and frequency at the same time, at least for your drive. Because the motor is inductive to keep currents reasonable as you reduce frequency you should reduce voltage V = L*di/dt, as MrAI stated.
To reduce the amplitude you just reduce the maximum PWM on duty cycle as you mentioned.
 
Thank you for the info, for snubbing why don't you just have fast diodes across each device? As long as the DC supply is well decoupled by metal film caps near the devices that should deal with your ringing. But it is very layout dependent.
The common core for your supplies shouldn't be a problem since the windings are isolated. But a spike in one supply will typically be coupled to all.
Soft start is to ramp voltage and frequency at the same time, at least for your drive. Because the motor is inductive to keep currents reasonable as you reduce frequency you should reduce voltage V = L*di/dt, as MrAI stated.
To reduce the amplitude you just reduce the maximum PWM on duty cycle as you mentioned.

Hi,

What do you mean by 'have fast diodes across each device' ?
Normally the diodes have to be reverse connected, so they cant snub any overshoot. Snubber diodes are connected so that they conduct when the voltage overshoots and that allows the snubber cap to absorb some of that energy and later the bleeder resistor (or other means) to dissipate it out.
 
MrAL and all other members thanks you explained well...I am gonna implement the way MrAL told....it seems much better than varying both the frequency and amplitude at the same time..as varying frequency sometimes disturbs the current cycle of sine.
Now if you please throw some light on snubber as well then it will be a complete cake with sugar....thanks :)
 
Hi,

What exactly do you need to know about the snubber?

The bleeder resistor has to be able to take the stored charge down enough between each switch cycle so that the voltage is not allowed to build up to a value greater than the rating of the transistor or else it does not snub well enough and the transistor will blow. We could go over come calculations, but i think this might involve some real time run experiments because the energy in the overshoot is difficult to determine theoretically. There are many factors that determine how much energy will need to be dissipated, such as rise time and circuit inductance in various places, so basically you try a resistor value and run it up little by little and watch the behavior. If the voltage rises too much you'll have to decrease the resistance of the snubber, and also increase the power rating most likely.
Also, the snubber circuits must be located physically very close to the device they are snubbing to minimize lead inductance which acts in a way that reduces the effectiveness of the snubber. For example if the transistor is mounted on a heat sink then the snubber should also be mounted on the same heatsink either on a small PC board or perhaps on ceramic standoffs using point to point wiring.
 
Better ac drives have 3 hefty chokes at the o/p, emi is a consideration for production quantities.
 
Reverse biased fast diodes will conduct the ring into either the V+ supply or ground. The slower bulk diodes are not fast enough for this. Assumes that the V+ and ground are well decoupled.
 
Thankyou I will give it a go..
Stay tuned....:)
There are more questions to come as I have seen maybe encoder or generator at the top of the motor will check it tomorrow.
Means its a closed loop.
Thanks everyone
 
Better ac drives have 3 hefty chokes at the o/p, emi is a consideration for production quantities.
Yes, I mentioned the same in one of my earlier posts. It also helps to limit di/dt in case of a motor fault condition giving your current protection time to protect.
 
Last edited:
Reverse biased fast diodes will conduct the ring into either the V+ supply or ground. The slower bulk diodes are not fast enough for this. Assumes that the V+ and ground are well decoupled.

Hi,

Thanks for explaining your original point, i understand what you mean better now.

What you are saying it seems is that a diode across both the upper and lower transistor will clamp the output of a half bridge to either ground or V+, and that is reasonable in theory. But in practice it is often not good enough because along with the assumption that the supply line is well decoupled there is also an assumption that the feed line to the transistor is very short, and that is often not the case in higher powered converters. The lead from the power supply to the collector or drain of the upper transistor for example could be somewhat long because it has to go from filter caps to the heatsink where the transistor is located. This lead naturally has inductance, and that is what the real problem is, not just the decoupling. As the voltage rises at the emitter or source, the diode conducts, but the inductance between the V+ supply and the diode prevents current flow for a short time, and unfortunately that short time even though it's very short is still too long to adequately clamp the voltage across the transistor. This is what ushers in the use of snubber circuits that are specially designed with very short leads so that they can act effectively on that short time period. Remember even a microsecond of overshoot could hurt the transistor.
 
Thank you for the info, for snubbing why don't you just have fast diodes across each device? As long as the DC supply is well decoupled by metal film caps near the devices that should deal with your ringing. But it is very layout dependent.
Already mentioned that it is layout dependent, but thanks for elaborating. You have to design it so lead lengths are short etc. But you have to do the same with a good snubber. 6 of 1 half a dozen of another.
 
ThAnkyou for explanation.
Can anyone please tell me how to calculate values I can provide the ringing waveform data ....
I build one but the wire length was too long now I am going to build another with short wire mosfet or directly on the mosfet as you guys are telling.
I will try to meaasure ringing freq tomorrow further more..is ringing load dependant?
If I build one snubber on low power motor will ringing time period vary on larger motor or is it transistor dependant to some extent?
 
I'm joining this conversation a bit late, however on a switcher there are 2 'rings', one when the switching device switches off, this is a higher freq usually and is caused by the leakage inductance of the transformer primary and associated wiring, and the parasitic capacitance of the switching device acting as a tank circuit, and the second is caused by the primary inductance ringing with the parasitic capacitance of the switching device when the energy in the primary is exhausted also acting as a tank circuit (this doesnt happen if the switcher is continuous).

Leakage inductance is largely affected by the transformer wind, a neat winding is good, also put the highest current wind nearest the core.

Switching device parasitic capacitance does vary a little, the devices datasheet tells you what you need.
Ringing freq will be affected by the load to an extent, depends on the topology.
 
Last edited:
Hi,
After starting my motor in the light of above discussion.
There is need of braking system.
I had not dealt with such high currents before.
There for I am concious before taking any step.
Kindly guid me about the breaking technique for my motor.
Starting current is 30amp at 20Hz.
35vdc while motor is 30v rms time we didn't have 48v48vdcvdc such high current so we applied 35vdc current will further drop at 48vdc
..

Can I swap any two phases two generate reverse torque to break the motor or should I add soft stop as well in my code...
Thanks..
 
Can I swap any two phases two generate reverse torque to break the motor or should I add soft stop as well in my code...
Thanks..
I am no expert in motor control, but I´d guess that if you just simply swap phases things will explode and burn (and / or shafts and gears will break...).
I think the correct approach would be to first stop driving the motor, then reverse the direction and start ramping the soft start up to gradually apply brake. Remeber that a fast spinning motor will require much lighter load than a slowly spinning one to achieve the same torque, so you need to start from zero current and gradually build up.
 
Yes it seems to be a gud solution will try it ...but what if truck is moving donw a hill in that case is this type of braking safe as motor will never come to rest except applying mechanical brake?
 
What do you mean by safe? Safe for the motor, or controller, or passengers?
Normally electrodynamic braking cannot make a complete stop, but if you can apply power to the motor in opposite direction than it is moving then you should have no trouble gradually stopping and then going into reverse. If you wanted to stop and stay still, then you would have to employ some kind of motion detection and apply just enough torque to cancel the gravity, but this might be not very good for the motors. Thus all electric vehicles like trams and metro have mechanical parking brakes.
 
Ok fine I m will give it a go ..
One more thing is there any diff if for stopping motor I just gradually decrease my amplitude of resultant lets say at 40hz rmsrmsrms 25volts then I gradually decrease this voltage to zero...
Or should I decrease frequency as well...?
 
Hi,
Can Anyone Please Tell Me Why I am Seeing This Waveform At Gate Of My FET Sorry For My Laziness I Painted My Circuit And Waveform.
What Could be the possible reason i have altered my gate resistor even shorted my gate to direct output of HCPL3120 Optocoupler but sudden shutdown peak is there during ON period it went off if i use a large value resistor but my On Time increases too much i can't afford..
Kindly Guide me in this regard.
 

Attachments

  • Untitled.png
    Untitled.png
    17.9 KB · Views: 128
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top